GO concurrency

Experiment on how to handle concurrency in real projects.

Suppose that, for each request, we have to access N times an upstream that can take too long to answer (s3 for example).

We need:

• Each work upstream request should be done in its own goroutine
• We should have a timeout, in case an upstream takes too long to return
• Collect the result of each upstream

Run

go run ./...

Test

go test -v ./...